PDZD8 overexpression alleviates early brain injury via regulating MERCs integrity after experimental subarachnoid hemorrhage

Abstract

Subarachnoid hemorrhage (SAH) is a devastating cerebrovascular disorder with high acute mortality and long-term neurological disability, and early brain injury (EBI) characterized by mitochondrial dysfunction, oxidative stress, and neuronal apoptosis is a pivotal determinant of poor prognosis. Mitochondria-endoplasmic reticulum contact sites (MERCs) are specialized membrane domains essential for maintaining cellular homeostasis via calcium trafficking and lipid exchange, but their regulatory mechanisms in SAH-induced EBI remain largely undefined. Here, we investigated the role and underlying mechanism of PDZD8, a core MERCs-stabilizing protein, in SAH pathogenesis using in vivo endovascular perforation models of male C57BL/6 mice and in vitro oxyhemoglobin (OxyHb)-challenged primary cortical neurons/HT22 cells, combined with PDZD8 overexpression, CRISPR/Cas9-mediated knockout, and C884A site-directed mutagenesis. Results demonstrated that PDZD8 was neuron-specifically downregulated at 48 h post-SAH, which closely correlated with MERCs structural disruption detected by transmission electron microscopy, impaired mitochondrial respiration analyzed via Seahorse assays, excessive reactive oxygen species production, and severe neuronal damage assessed by Nissl staining. PDZD8 overexpression preserved MERCs integrity, restored mitochondrial metabolic balance, mitigated oxidative stress, and ameliorated neurobehavioral deficits evaluated by modified neurological severity scores, rotarod, and open field tests. Mechanistically, SAH-induced PDZD8 downregulation was associated with enhanced global S-nitrosylation and post-translational regulation at cysteine 884 (C884), promoting its ubiquitination and proteasomal degradation, while C884A mutation abrogated this process. Our findings reveal a previously unrecognized mechanism involving S-nitrosylation-associated ubiquitination of PDZD8 in SAH-induced MERCs dysfunction and EBI, highlighting PDZD8 as a promising therapeutic target for SAH treatment.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12974-026-03931-y.

Introduction

Subarachnoid hemorrhage (SAH) is a complex and devastating cerebrovascular disorder characterized by the rupture of cerebral vessels and blood extravasation into the subarachnoid space, accounting for approximately 10% of all stroke cases worldwide [1]. In China, the incidence of SAH is estimated at 1–2 per 100,000 person-years, with a mortality rate as high as 50% within 72 h of onset, making it one of the most life-threatening neurological emergencies [2]. A growing body of evidence indicates that early brain injury (EBI), defined as acute cerebral damage occurring within 72 h post-SAH, is a pivotal determinant of poor prognosis, including high mortality and long-term neurological deficits [3]. The pathological mechanisms underlying EBI are highly complex, involving a cascade of events such as elevated intracranial pressure, neuroinflammation, oxidative stress, blood-brain barrier disruption, and acute neuronal death [4, 5]. Despite advances in clinical management, the therapeutic options for mitigating EBI remain limited, highlighting an urgent need to identify novel molecular targets and protective mechanisms.

Mitochondria play a central role in maintaining neuronal survival and function, as they are responsible for energy production (ATP synthesis), calcium homeostasis regulation, reactive oxygen species (ROS) metabolism, and participation in neurodevelopment [6]. In the context of SAH, mitochondrial dysfunction emerges as a key contributor to EBI: pathological processes such as oxidative stress and neuroinflammation disrupt mitochondrial integrity, leading to impaired energy metabolism, excessive ROS production, and dysregulated calcium signaling, which ultimately trigger neuronal apoptosis or necrosis [7, 8]. Mitochondrial homeostasis is tightly regulated by multiple quality control mechanisms, including mitochondrial autophagy, dynamics (fission and fusion), and protein quality control [9]. Notably, accumulating experimental evidence suggests that preserving mitochondrial homeostasis may serve as a promising therapeutic strategy for alleviating EBI post-SAH [10–12], underscoring the need to explore the molecular pathways governing mitochondrial function in this pathological setting.

A critical mediator of mitochondrial homeostasis is the mitochondria-endoplasmic reticulum (ER) contact sites (MERCs), specialized membrane domains formed by physical proximity between mitochondria and the ER [13]. These structures are not static but dynamically regulate the distance between the two organelles, enabling efficient communication and material exchange, including calcium transfer, lipid metabolism, and ROS signaling [14].

Emerging studies in stroke models have highlighted the significance of MERCs in neuroprotection: in a mouse middle cerebral artery occlusion model, MERCs structure was significantly reduced post-ischemia [15]; overexpression of key MERCs-forming proteins promoted axonal regeneration after neural injury [16], while knockout of these proteins led to excessive mitophagy, increased ROS production, and expanded infarct volumes [15]. Given these findings, it is reasonable to hypothesize that MERCs disruption may contribute to mitochondrial dysfunction and EBI post-SAH, and that preserving MERCs integrity could be a viable approach to maintain mitochondrial homeostasis and improve outcomes.

PDZ domain containing 8 (PDZD8) is established as a key structural component of MERCs [17]. Structurally, PDZD8 contains eight PDZ domains, which mediate protein-protein interactions, and a sterile alpha motif domain that enables it to function as an ER membrane protein and dynamically bind to mitochondrial lipid membranes, thereby stabilizing MERCs integrity [18, 19]. Existing research on PDZD8 has focused on its roles in neuronal calcium exchange, lipid metabolism, and mitophagy [20, 21], but its expression pattern, functional role, and regulatory mechanisms in SAH-induced EBI remain completely unexplored. Given its essential role in maintaining MERCs structure, we speculate that PDZD8 may be involved in regulating mitochondrial homeostasis post-SAH by preserving MERCs integrity, representing a novel molecular target for EBI intervention.

In summary, the present study aimed to investigate the role and mechanism of PDZD8 in regulating MERCs structure and mitigating EBI post-SAH. We hypothesized that SAH leads to downregulation of PDZD8, resulting in MERCs disruption, mitochondrial dysfunction, and subsequent neuronal injury; conversely, upregulation of PDZD8 preserves MERCs integrity, maintains mitochondrial homeostasis, and thereby exerts neuroprotective effects against EBI.

To validate this hypothesis, we employed in vivo SAH models (endovascular perforation) and in vitro oxyhemoglobin (OxyHb)-induced neuronal injury models, combined with PDZD8 overexpression/knockout strategies. This research provides new insights into the molecular mechanisms of EBI and identifies potential therapeutic strategies for SAH.

Methods and materials

Experimental animals and in vivo establishment of the SAH model

The SAH model was established via the endovascular perforation method, as documented previously [22]. Adult male C57BL/6 mice (21–28 g, 10–12 weeks old) were procured from the Animal Center of the Chinese Academy of Sciences (Shanghai, China). This study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (Approval No.: GSWS2023022), complying with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and ARRIVE guidelines [23]. Mice were housed in a controlled environment (22–24 °C, 50–60% relative humidity, 12 h light-dark cycle) with free access to food and water. Computer-generated randomization and double-blind procedures were applied to minimize bias.

For modeling, mice were anesthetized with 2% isoflurane. A sharpened 6–0 nylon suture was inserted into the right internal carotid artery via the external carotid artery stump, advanced 2–3 mm after resistance was felt, then withdrawn. The incision was sutured, and mice were observed for 1 h postoperatively with body temperature maintained. At sacrifice, SAH severity was graded; mice with scores ≥ 8 were included, while those < 8 were excluded.

The sham group underwent identical procedures without arterial perforation. Sham controls were collected at time-matched intervals corresponding to each experimental group. Mortality and exclusion rates are shown in Table S1.

Experimental design

Five sets of experiments were designed to explore the role and mechanism of PDZD8 in SAH and OxyHb-induced neuronal injury.

PDZD8 time-course, localization and MERCs changes post-SAH

Mice were assigned to 3 h, 6 h, 12 h, 24 h, 48 h, and 72 h post-SAH groups (n=6/group). Ventral temporal cortical tissues were harvested for Western blotting (WB, PDZD8 and global S-nitrosylation), quantitative reverse transcription polymerase chain reaction (qRT-PCR, PDZD8 mRNA), immunofluorescence (PDZD8 localization), and transmission electron microscopy (TEM, MERCs integrity).

PDZD8 overexpression alleviates SAH-induced MERCs and neuronal damage

Mice were randomized into 6 groups (n = 6/group). WB and immunofluorescence verified PDZD8 overexpression; TEM evaluated MERCs structure and Nissl staining assessed neuronal injury.

PDZD8 overexpression improves post-SAH neurobehavior

70 mice survived to 21 days post-SAH, divided into 6 groups (n=12,12,12,11,11,12). Modified neurological severity scores (mNSS), Rotarod and open field tests (OFT) assessed neurobehavioral functions.

PDZD8 overexpression mitigates OxyHb-induced injury in primary neurons

Neurons were divided into 6 groups (n=3/group). WB and immunofluorescence verified PDZD8 overexpression; proximity ligation assay (PLA) detected MERCs formation; MitoSOX, oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) evaluated mitochondrial function.

Mechanism of PDZD8 reduction in OxyHb-treated HT22 cells

Wild-type and PDZD8 knockout (PDZD8 KO) cells were used for rescue experiments with wild-type and point-mutated plasmids. WB detected total and S-nitrosylated (SNO) PDZD8; co-immunoprecipitation examined PDZD8-ubiquitin interaction (details in Supplementary Fig. 1).

Cell culture and OxyHb intervention model

Primary cortical neurons were cultured as previously described [24]. Briefly, cortical tissues from fetal mouse brains (meninges and blood vessels excised) were digested with 0.25% trypsin-EDTA at 37 °C for 5 min, rinsed three times with PBS to terminate digestion, and centrifuged at 1500 rpm for 5 min. The cell pellet was resuspended in neurobasal medium supplemented with 2% B27, 2 mmol/L L-glutamine, 50 U/mL penicillin, and 50 µg/mL streptomycin (all from Gibco, Carlsbad, CA, USA), then seeded into 6-well or 12-well plates with fresh medium; half the medium was replaced every 48 h. HT22 cells (a mouse hippocampal neuronal cell line, Shanghai Zhong Qiao Xin Zhou Biotechnology Co., Ltd.) were cultured in DMEM/F12 (Gibco, USA) containing 10% fetal bovine serum and 100 U/mL penicillin-streptomycin, maintained at 37 °C in a 5% CO₂ incubator.

The in vitro SAH model was induced via OxyHb stimulation [22, 24]. OxyHb was prepared by lyophilizing freeze-dried mouse hemoglobin (MP Biomedicals LLC, USA) per the manufacturer’s instructions. Neurons were exposed to 10 µM OxyHb at 37 °C in 5% CO₂ for 6 h; the medium was then discarded, followed by three PBS washes. For in vitro PDZD8 overexpression assays, cells were pretreated with plasmids for 3 days prior to OxyHb exposure, following the manufacturer’s protocol.

In vivo lentivirus transfection

Lentivirus for PDZD8 overexpression and negative control CON254 (Genechem Co., Shanghai, China) were stored at -80 °C. For intracerebroventricular injection, the lentivirus was diluted to 4 × 10⁸ TU/mL in enhanced transfection solution. Stereotactic injection was performed into the right lateral ventricle of mice, using coordinates relative to bregma: (1) anterior-posterior: 0.37 mm from bregma, medio-lateral: 0.75 mm in the right hemisphere, dorso-ventral: 2.5 mm below the dura mater. Each site was injected with 0.5 µL of lentivirus at a rate of 0.2 µL/min. The needle was retained for 5 min after injection. The SAH model was established 7 days later. Detailed lentivirus information is provided in Supplementary Table S2.

Plasmid transfection

PDZD8 plasmids were constructed based on the cytomegalovirus enhancer-multiple cloning site-3FLAG-SV40-Puromycin vector system. Detailed information regarding plasmid sequences and cloning protocols is provided in Supplementary Table S3-S4. HT22 cells were transfected using Lipofectamine 3000 (Invitrogen, NY, USA) in strict accordance with the manufacturer’s instructions. Specifically, cells seeded in 6-well plates were subjected to transfection when reaching 50%–60% confluence. A transfection mixture containing 2.5 µg of plasmid DNA and 5 µL of Lipofectamine 3000 was prepared in Opti-MEM medium, followed by a 15-minute incubation at room temperature prior to being added to the cells. The PDZD8-expressing plasmids were delivered to the OxyHb + OE-PDZD8 group, whereas the empty vector Con520 was used as the negative control.

Generation of PDZD8-KO HT22 cell line by CRISPR/Cas9

For PDZD8 knockout, two single-guide RNAs carrying protospacer-adjacent motifs of CGG and GGG, respectively, were designed for targeting the exon E1 region of the PDZD8 gene.

The sgRNA design was performed using the online CRISPR design platform Red Cotton™ (Guangzhou, China; https://en.rc-crispr.com/). Oligonucleotide pairs corresponding to the two targeting sites were annealed and subcloned into the YKO-RP006 vector (Ubigene Biosciences Co., Ltd., Guangzhou, China). The recombinant plasmids encoding Cas9 nuclease and the two sgRNA sequences were transfected into HT22 cells via electroporation with Lipofectamine 3000 (Invitrogen, NY, USA). At 48 h post-transfection, puromycin was supplemented into the culture medium for stable cell line selection. Subsequently, single-cell clones were isolated via the limiting dilution method and seeded into 96-well plates. Putative PDZD8-knockout (PDZD8-KO) cell clones were verified through PCR amplification and Sanger sequencing of genomic DNA (Fig. S2). The sgRNA sequences and validation primers utilized for CRISPR/Cas9-mediated PDZD8 editing are listed as follows: g1: GTTCGTGAAGCTGTCTCGCG TGG, g2: AGGTGCGCTCCCAGTTCGAG GGG, F1: GCTGCTCCTGATCCTGGCGTC, R1: CACGGGAAGTGCCCACTACCC, F2: CGTGGAGGACCCGCTGATTGAC.

Western blot

Proteins from cells and tissues were extracted using lysis buffer (Beyotime, China). Protein concentrations were determined via the BCA protein assay kit (Beyotime, China). A total of 20 µg protein (10 µl of 2 µg/µl sample) was loaded per well. Proteins were separated by SDS-PAGE (100 V, 90 min) on a Bio-Rad Mini-PROTEAN Tetra system, using 6%, 10%, or 12% resolving gels (selected based on target protein molecular weight) and 5% stacking gels.

Subsequently, proteins were transferred onto nitrocellulose membranes (HATF00010; Millipore Sigma, USA) via the Bio-Rad Trans-Blot Turbo system under wet transfer conditions (250 mA, 90 min). Membranes were blocked with 5% BSA for 1 h, then incubated overnight at 4 °C with primary antibodies (Table S5). After washing, HRP-conjugated secondary antibodies were applied; details of secondary antibodies and their dilutions are provided in Supplementary Table S6.

Protein bands were visualized using ECL substrate with a Chemiscope 5300 imager (Clinx, China). Relative protein quantification was performed using ImageJ software (National Institutes of Health, USA) and normalized to β-tubulin.

Detection of SNO proteins

SNO proteins were assayed with the Thermo Scientific Pierce S-Nitrosylation Western Blot Kit (Cat. No. 90105). Briefly, 100–200 µg of total protein per sample was extracted using HENS Buffer and quantitated by BCA assay. Free cysteine thiols were blocked via incubation with 20 mM methyl methanethiosulfonate for 30 min at room temperature, after which excess methyl methanethiosulfonate was removed by acetone precipitation. The resuspended protein pellets were treated with sodium ascorbate to selectively reduce S-nitrosocysteines, followed by labeling with iodoTMT reagent for 1–2 h. Proteins were separated by SDS-PAGE, transferred onto nitrocellulose membranes, and blocked with 5% nonfat dry milk. Membranes were probed sequentially with anti-TMT primary antibody (1:1000) and HRP-conjugated secondary antibody, then visualized using ECL substrate and an imager. SNO-PDZD8 was specifically detected with an anti-PDZD8 antibody, and its relative level was calculated as the ratio of SNO-PDZD8 to total PDZD8.

Immunofluorescence analysis

Brain tissues were fixed with 4% paraformaldehyde, paraffin-embedded, and cut into 4 μm-thick sections. Deparaffinization was achieved with xylene and gradient ethanol, followed by antigen retrieval in citrate buffer via heat treatment. Permeabilization was performed by incubating sections with 0.5% Triton X-100 for the indicated time. Subsequently, sections were blocked with 5% BSA for 1 h at room temperature to eliminate non-specific binding.

Primary antibodies were incubated with sections overnight at 4 °C, followed by secondary antibody incubation for 1 h at room temperature. Fluorescence images were acquired with a Nikon ECLIPSE Ni-U/DS-Ri2 microscope (Nikon, Japan).

Proximity ligation assay

PLA was conducted with the Duolink In Situ Red Starter Kit (Mouse/Rabbit; Sigma-Aldrich, Cat. No. DUO92101) following the manufacturer’s protocols. Briefly[25], mouse primary neurons were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100.

Cells were then incubated overnight at 4 °C with primary antibodies in blocking buffer (PBS containing 0.1% Triton X-100 and 4% BSA). Oligonucleotide-conjugated secondary antibodies (Duolink PLUS anti-rabbit IgG and MINUS anti-mouse IgG) were applied at a 1:5 dilution in blocking buffer and incubated for 1 h at 37 °C. Subsequently, ligation solution (1:5 diluted ligation buffer and 1:40 diluted ligase in ultrapure water) was incubated for 30 min at 37 °C, followed by incubation with amplification solution (1:5 diluted amplification buffer and 1:40 diluted polymerase in ultrapure water) for 1 h 40 min at 37 °C. After the reactions, slides were mounted using Duolink In Situ Mounting Medium containing DAPI.

Transmission electron microscopy

Samples were pre-fixed in 2.5% glutaraldehyde (0.1 M phosphate buffer) for 2 h at room temperature, washed three times with 0.1 M phosphate buffer (15 min each), then post-fixed in 1% osmium tetroxide for 2 h. After another three washes, gradient dehydration was performed with ethanol and acetone (15 min each). Samples were infiltrated with SPI 812 embedding medium, embedded, and polymerized at 37 °C and 60 °C sequentially. Ultrathin Sects.  (60–80 nm) were cut with a Leica UC7 ultramicrotome, stained with 2% uranyl acetate (30 min) and lead citrate (15 min), and imaged using a JEM 1400. Quantitative morphometric analysis was performed using ImageJ software. Mitochondrial perimeter, the distance between ER and mitochondria, and the percentage of mitochondrial perimeter in contact with ER were directly measured from TEM images. Mitochondrial roundness index and the ER-mitochondria contact coefficient (ERMICC) was determined according to established protocols [25, 26].

Co-Immunoprecipitation

Co-Immunoprecipitation was performed as follows: total proteins were incubated with primary antibodies (see Table S5) under rotation at 4 °C overnight. Subsequently, samples were mixed with 100 µL pre-treated protein A/G agarose beads (B23202; Selleck, USA) and incubated for an additional 1 h at 4 °C on a rocker. Immune complexes were isolated via magnetic separation, with the supernatant discarded. The complexes were washed five times with precooled lysate (without protease inhibitors) to eliminate unbound proteins. Next, 1× loading buffer was added, and samples were heated at 100 °C for 10 min to elute complexes from the beads. Samples were ultimately analyzed by Western blot.

Quantitative reverse transcription polymerase chain reaction for mRNA

Total RNA was extracted from brain tissues using the Cell/Tissue Total RNA Rapid Extraction Kit (Cat. No. M5105; NCM Bio, Suzhou, China) following the manufacturer’s instructions. cDNA was synthesized from 1 µg of total RNA using the One-Step High-Efficiency RT Kit (Cat. No. M7100; NCM Bio) in a 20 µL system: 13 µL RNA/DEPC-treated H₂O mixture, 4 µL 5×RT Mix, and 3 µL All-in-One Enzyme Mix. Reaction conditions: 37 °C for 2 min (gDNA removal), 55 °C for 15 min (RT), 85 °C for 5 min (enzyme inactivation).

qRT-PCR was performed with SYBR Green Premix in 20 µL systems containing 10 µL Premix, 2 µL diluted cDNA (1:10), and 0.3 µM each of primers (PDZD8 forward: GCTCATTGCTATTGGAGGTGTG, PDZD8 reverse: AGCTTTCTTCCAACTGGCCC; GAPDH forward: GGAAATCCCATCACCATCTTC, GAPDH reverse: TGGACTCCACGACGTACTCAG, GAPDH as internal control). Cycling conditions: 95 °C for 30 s, 40 cycles of 95 °C/5 s and 60 °C/30 s. Melt curve analysis verified specificity, and relative expression was calculated via 2^-ΔΔCt by blinded investigators (triplicate assays per sample).

Nissl staining

Nissl staining was performed to assess neuronal morphological changes in the cerebral cortex post-SAH. Dewaxed paraffin-embedded brain sections were rinsed three times with distilled water, then incubated with Nissl stain at 55 °C for 40 min (solution air-dried during incubation) followed by three additional distilled water rinses. Sections were sequentially immersed in gradient ethanol and xylene for decolorization and clearing, then mounted with neutral balsam. Nissl-positive cell counts in the temporal cortex was conducted at 400× magnification using a Nikon ECLIPSE Ni-U/DS-Ri2 optical microscope (Nikon, Kanagawa, Japan).

Measurement of OCR and ECAR

OCR and ECAR were measured using the Seahorse XF24 Extracellular Flux Analyzer (Agilent Technologies, USA) to evaluate mitochondrial respiration and glycolytic metabolism. Briefly, cells were seeded into XF24 microplates at a density of 5 × 10³ cells/well and cultured for 48 h in DMEM supplemented with 10% FBS, 25 mM glucose, and 2 mM glutamine at 37 °C in 5% CO₂. Before the assay, the XF sensor cartridge was hydrated and calibrated overnight according to the manufacturer’s instructions. On the day of measurement, culture medium was replaced with Seahorse XF assay medium and cells were equilibrated at 37 °C in a non-CO₂ incubator for 30 min.

For OCR measurement, mitochondrial respiration was assessed using the Seahorse XF Cell Mito Stress Test protocol. Baseline OCR was recorded followed by sequential injections of oligomycin, FCCP, and rotenone/antimycin A. Basal respiration, ATP production, maximal respiration, and spare respiratory capacity were calculated.

For ECAR measurement, glycolytic function was evaluated using the Seahorse XF Glycolysis Stress Test protocol. Baseline ECAR was recorded followed by sequential injections of glucose, oligomycin, and 2-deoxy-D-glucose (2-DG). Glycolysis and glycolytic capacity were calculated.

Measurements were recorded every 10 min with three measurement cycles after each injection. All experiments were independently repeated three times.

Neurobehavioral assessment

Modified neurological severity scores

mNSS was used to assess mice’s sensory, motor, balance, and reflex functions as previously described. The test includes 10 items, with scores ranging from 0 (normal) to 18 (severe impairment); 1 point was awarded for failed tasks or absent responses. Mice were trained preoperatively to ensure a baseline score of 0, and blind scoring was performed on days 1, 3, 5, and 7 post-SAH.

Rotarod test

The rotarod test was conducted using a SA102 rotator fatigue instrument (SANS Biotechnology, China). Mice were placed on a rotating drum with acceleration from 4 to 40 RPM over 5 min, and the latency to fall was recorded.

Open field test

OFT evaluated exploratory activity and anxiety-like behavior 21 days post-SAH. Mice were individually placed in the center of a 95 × 95 × 95 cm wooden arena (divided into 16 sections) for 10-min free exploration. Activity was captured by an overhead camera, and total length traveled, immobility time, and straddle lattice times were analyzed via video tracking. The arena was cleaned with 75% ethanol after each trial.

Statistical analysis

Statistical analyses were conducted with GraphPad Prism 9.0, and results were presented as mean ± standard deviation (SD). Normality was tested via the Shapiro-Wilk test, and variance homogeneity via the Brown-Forsythe test. Pairwise group comparisons were analyzed using two-tailed unpaired Student’s t-test. One-way analysis of variance (ANOVA) with Bonferroni post hoc test and two-way ANOVA with Dunnett’s post hoc test were applied for multiple group comparisons and repeated measures, respectively. Statistical significance was defined as p < 0.05.

Results

PDZD8 is downregulated in neurons post-SAH and correlates with MERCs disruption

A representative Western blot image of PDZD8 is shown in Fig. 1A. Time-course Western blot analysis was performed to track PDZD8 expression dynamics, revealing that PDZD8 protein levels in the ventral temporal cortex showed a gradual downward trend at 12 h and 24 h post-SAH but without statistical significance, and were significantly reduced at 48 h post-SAH compared with the sham group (Fig. 1B). Immunofluorescence staining showed that PDZD8 signal was predominantly enriched in NeuN-positive neurons (Fig. 1C). Quantitative analysis further demonstrated a significant reduction in neuronal PDZD8 signal at 48 h post-SAH, consistent with the Western blot results (Fig. 1D). In contrast, limited colocalization of PDZD8 signal was observed with the astrocyte marker GFAP in both sham and SAH 48 h groups (Fig. 1E), and similarly limited colocalization was detected with the microglial marker IBA1 (Fig. 1F). These findings suggest that PDZD8 expression was primarily enriched in neuronal populations under the conditions examined.

Fig. 1.

PDZD8 is neuron-specific, downregulated post-SAH, and associated with MERCs disruption. A Representative Western blot image showing PDZD8 expression in the ventral temporal cortex. B Quantitative analysis of PDZD8 protein levels at different time points post-SAH. C Immunofluorescence images of PDZD8 (green) and neuronal marker NeuN (red); DAPI (blue) labels nuclei (scale bar=40 μm). D Quantitative analysis of PDZD8 fluorescence intensity (n=6 per group). Fluorescence intensity was quantified in NeuN-positive cells. E Immunofluorescence staining for PDZD8 (green) and GFAP (red) in sham and SAH 48 h groups; scale bar=40 μm. F Immunofluorescence staining for PDZD8 (green) and IBA1 (red) in sham and SAH 48 h groups; scale bar=40 μm. No colocalization observed in E,F. G Representative transmission electron microscopy images (two rows × three columns): top row (sham group), bottom row (SAH 48h group); middle column (scale bar=2 μm), left/right columns (magnified regions); red solid lines indicate mitochondria-endoplasmic reticulum contact sites, red dashed lines indicate their distance. H–M Quantitative TEM analysis: (H) ER–mitochondria distance; (I) percentage of mitochondrial contact with ER; (J) percentage of ER contact with mitochondria; (K) mitochondrial roundness index; (L) mitochondrial perimeter; (M) ER–mitochondria contact coefficient (ERMICC). (n=30 cells). Data are presented as mean ± SD; n.s., not significant, **P ≤ 0.01 vs. sham group. Statistical analyses were performed using the methods described in Methods

TEM images (Fig. 1G) illustrated obvious morphological alterations of MERCs in the SAH 48 h group. Specifically, the distance between endoplasmic reticulum and mitochondria was notably elevated (Fig. 1H), while the percentages of contact regions on both mitochondria and endoplasmic reticulum were markedly reduced compared with the sham group (Fig. 1I, J). Quantitative analysis of mitochondrial morphological parameters showed no statistically significant differences in mitochondrial roundness index (Fig. 1K) or perimeter (Fig. 1L) between the sham and SAH 48 h groups, indicating that SAH does not alter global mitochondrial size or shape at this time point. In contrast, the ERMICC, a comprehensive metric reflecting MERCs structural integrity, was significantly decreased in the SAH 48 h group compared with the sham group (Fig. 1M). These cumulative findings collectively demonstrate that PDZD8 is a neuron-specific protein whose expression is prominently downregulated at the 48 h time point post-SAH, and this significant reduction in PDZD8 levels is closely associated with substantial structural disruption of MERCs.

PDZD8 overexpression alleviates SAH-induced MERCs disruption and neuronal injury

To explore whether PDZD8 overexpression alleviates SAH-induced MERCs disruption and neuronal damage, we constructed PDZD8-overexpressing lentivirus carrying a FLAG tag. Western blot analysis (Fig. 2A) showed the FLAG tag was exclusively detected in PDZD8 overexpression groups, and PDZD8 protein levels were significantly elevated in both sham and SAH 48 h overexpression groups compared with their respective vector groups (Fig. 2B). Immunofluorescence staining of PDZD8 and neuronal marker NeuN further confirmed that lentivirus-mediated PDZD8 overexpression notably increased PDZD8 levels in neurons, consistent with Western blot results (Fig. 2C and D). These findings validate successful PDZD8 overexpression in neuronal populations of both sham and SAH models.

Fig. 2.

Validation of PDZD8 overexpression in sham and SAH models. A Representative Western blot images showing PDZD8 and FLAG tag expression; FLAG was only detectable in PDZD8 overexpression groups. B Densitometric quantification of PDZD8 protein levels (n=6 per group). C Immunofluorescence staining for PDZD8 (green), neuronal marker NeuN (red), and DAPI (blue, nuclear staining; scale bar=40 μm). D Quantification of PDZD8 fluorescence intensity in neurons (n=6 per group). Fluorescence intensity was quantified specifically in NeuN-positive cells. Statistical comparisons were performed between the following groups: sham vs SAH 48 h and SAH 48 h+Vector vs SAH 48 h+OE-PDZD8. Data are presented as mean ± SD; **P ≤ 0.01. Statistical analyses were performed using the methods described in Methods

TEM analysis, as shown in Fig. 3A, revealed marked structural disruption of MERCs in SAH groups compared with sham groups, characterized by increased ER-mitochondria distance and reduced contact regions. In contrast, PDZD8 overexpression notably ameliorated these pathological alterations: the distance between ER and mitochondria was significantly shortened, and the percentages of mutual contact regions on both ER and mitochondria were notably increased in the SAH PDZD8 overexpression group relative to the SAH vector group, as quantified in Fig. 3B-D. To comprehensively quantify mitochondrial morphology and overall MERCs structural integrity across all six experimental groups, we further measured mitochondrial roundness index, mitochondrial perimeter, and ER-mitochondria contact coefficient (ERMICC) (Fig. 3E-G). Statistical comparisons were performed between two key groups: sham vs. SAH 48 h, and SAH 48 h+Vector vs. SAH 48 h + OE-PDZD8. No statistically significant differences were observed in either mitochondrial roundness index or perimeter in these two comparisons. In contrast, ERMICC was significantly decreased in the SAH 48 h group compared with the sham group, and PDZD8 overexpression significantly elevated ERMICC levels in SAH rats compared with the SAH 48 h+Vector group. Further TEM observations showed that dense aggregates within mitochondria were exclusively detected in SAH-related groups, and PDZD8 overexpression was associated with smaller-sized aggregates, suggesting a potential role in mitigating mitochondrial damage.

Fig. 3.

PDZD8 overexpression restores MERCs structure and reduces neuronal damage post-SAH. A Representative TEM images; red solid lines indicate ER-mitochondria contact sites, red dashed lines indicate their distance, red arrows point to dense aggregates in mitochondria (scale bar=1 μm). B–G Quantitative TEM analysis: (B) ER-mitochondria distance, (C) percentage of mitochondria in contact with ER, (D) percentage of ER in contact with mitochondria, (E) mitochondrial roundness index, (F) mitochondrial perimeter, (G) ER-mitochondria contact coefficient (ERMICC). H Representative Nissl staining images; red arrows indicate damaged neurons (scale bar=40 μm). I Quantitative analysis of Nissl-positive neuron counts per field. Data are presented as mean ± SD, n.s., not significant, *P ≤ 0.05, **P ≤ 0.01. Statistical analyses were performed using the methods described in Methods

In parallel with TEM findings, Nissl staining illustrated in Fig. 3H demonstrated that SAH led to a dramatic reduction in Nissl-positive neurons, reflecting severe neuronal injury.

However, PDZD8 overexpression significantly reversed this decrease, restoring the number of viable neurons. Notably, in sham groups, PDZD8 overexpression had no obvious impact on the count of Nissl-positive neurons, indicating the specificity of its protective effect in the pathological context of SAH (quantified in Fig. 3I). These results collectively demonstrate that PDZD8 overexpression rescues SAH-induced MERCs structural impairment and effectively mitigates neuronal injury, highlighting its neuroprotective role post-SAH.

PDZD8 overexpression ameliorates neurobehavioral deficits post-SAH

To evaluate the effect of PDZD8 overexpression on neurobehavioral function after SAH, mNSS, rotarod test, and OFT were performed with mice trained for 3 days pre-experiment, and all tests included 6 groups with no significant differences observed among sham subgroups at all time points. The mNSS was used to assess neurological deficits at 1, 3, 5, and 7 days post-SAH, with results shown in Fig. 4A. All groups scored 0 before SAH modeling; SAH induction led to significantly elevated mNSS scores, while PDZD8 overexpression resulted in scores notably lower than those of the SAH+vector group at each time point, indicating alleviated neurological impairment. The rotarod test (Fig. 4B) was conducted at 3, 7, and 14 days post-SAH to evaluate sensory-motor function. SAH mice exhibited shortened latency to fall off the rotarod compared with sham controls, whereas PDZD8 overexpression reversed this impairment across all tested time points relative to the SAH+vector group. The OFT was performed on day 21 post-SAH to assess exploratory activity and anxiety-like behavior. Figure 4C shows the OFT track heat maps (upper panel) and trajectory maps (lower panel), visualizing behavioral changes. Quantitative analyses (Fig. 4D–F) revealed that SAH induction resulted in decreased total movement distance (Fig. 4D), increased immobility time (Fig. 4E), and reduced straddle lattice times (Fig. 4F). PDZD8 overexpression significantly ameliorated these deficits compared with the SAH+vector group, as evidenced by increased total distance and straddle lattice times, as well as reduced immobility time. These results demonstrate that PDZD8 overexpression alleviates SAH-induced neurobehavioral impairments.

Fig. 4.

Effect of PDZD8 overexpression on SAH-induced neurobehavioral outcomes. A Modified neurological severity score (mNSS) of sham, sham + vector, sham+OE-PDZD8, SAH, SAH + vector, and SAH+OE-PDZD8 groups at 1, 3, 5, and 7 days post-SAH. B Fall latency in rotarod test (total duration: 300 s) of the six groups at 3, 7, and 14 days post-SAH. C Open field test (OFT) heat maps (upper panel), trajectory diagrams (lower panel) of the six groups at 21 days post-SAH (total duration: 600 s). D Total movement distance in OFT. E Immobility time in OFT. F Straddle lattice times in OFT. Data are presented as mean ± SD. **P ≤ 0.01; n = 11–12 for all subgroups in (A–F). Statistical analyses were performed using the methods described in Methods

PDZD8 overexpression alleviates SAH-mimicked neuronal injury via preserving MERCs integrity and mitochondrial homeostasis in vitro

To explore the mechanism underlying PDZD8-mediated neuroprotection against SAH-induced injury in vitro, we transfected primary mouse neurons with FLAG-tagged PDZD8 overexpression plasmids. Western blot assays (Fig. 5A) demonstrated that PDZD8 protein expression was markedly upregulated in both control and OxyHb-treated (SAH-mimicking) overexpression groups relative to their corresponding empty vector-transfected groups (Fig. 5B). Immunofluorescence co-staining of PDZD8 and NeuN further corroborated that plasmid-mediated transfection significantly elevated PDZD8 abundance in neurons (Fig. 5C, D).

Fig. 5.

PDZD8 overexpression rescues OxyHb-induced MERCs impairment in primary mouse neurons. A Western blot analysis of FLAG tag and PDZD8 protein expression in primary neurons transfected with empty vector or FLAG-tagged PDZD8 overexpression plasmid. B Quantitative analysis of PDZD8 protein levels in (A) (n=3 per group). C Immunofluorescence staining of PDZD8 (green) and neuronal marker NeuN (red) in primary neurons; scale bar = 40 μm. D Quantitative analysis of PDZD8 fluorescence intensity (n=3 per group). Fluorescence intensity was quantified in individual neurons. E Schematic diagram illustrating the principle of proximity ligation assay (PLA), with steps labeled as a, b, c, d, e: a. the primary antibody binds to two proteins of interest; b. PLA secondary antibody(plus and minus) combined with primary antibody; c. DNA linked to PLA probes to form ring structure; d. polymerase amplification of DNA loops; e. fluorescent probes bind to amplified DNA. F Representative PLA images showing IP3R-VDAC1 interactions (red puncta, indicative of MERCs) in the six groups (control, control + vector, control + PDZD8 OE, OxyHb 48h, OxyHb 48h + vector, OxyHb 48h + PDZD8 OE) of primary neurons; scale bar = 10 μm. G Quantitative analysis of IP3R-VDAC1 PLA puncta per cell (30 cells counted per group). Data are presented as mean ± SD. *P ≤ 0.05, **P ≤ 0.01. Statistical analyses were performed using the methods described in Methods. Figure 5E was created with BioRender.com

Subsequently, PLA targeting IP3R-VDAC1 interactions was performed to assess MERCs structure as previously described [25]. A simplified schematic of the PLA principle is shown in Fig. 5E; the specificity of this assay was rigorously validated through three standard negative control experiments, with representative images provided in Fig. S3. Representative images (Fig. 5F) revealed abundant MERCs puncta in the control group, with no obvious difference in the PDZD8 overexpression-only group. In contrast, OxyHb-treated neurons exhibited a marked reduction in MERCs puncta, whereas PDZD8 overexpression significantly elevated MERCs puncta number compared with the vector group in OxyHb-treated cells. Quantitative analysis (Fig. 5G) further verified these findings.

Seahorse assays simultaneously measured OCR and ECAR profiles (Fig. 6A), with no significant differences detected among the three control subgroups. OxyHb exposure induced a substantial decrease in basal respiration, ATP production, maximal respiration and spare respiratory capacity (Fig. 6B–E), whereas PDZD8 overexpression significantly ameliorated these declines compared with the OxyHb + vector subgroup. Conversely, OxyHb stimulation led to a marked elevation in ECAR-associated glycolysis and glycolytic capacity (Fig. 6F–G), effects that were significantly inhibited by PDZD8 overexpression. Representative MitoSOX Red and DAPI fluorescence images (Fig. 6H) and quantitative analysis (Fig. 6I) revealed no difference among control subgroups, while OxyHb increased positive cells and PDZD8 overexpression reversed this trend.

Fig. 6.

PDZD8 overexpression restores mitochondrial metabolic balance and alleviates oxidative stress in OxyHb-treated primary neurons. A Dynamic profiles of mitochondrial oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) in six experimental groups. B–E Quantitative analysis of OCR-linked mitochondrial respiratory parameters: basal respiration, ATP production, maximal respiration, and spare respiratory capacity; n = 3. F–G Quantitative analysis of ECAR-associated glycolytic parameters: glycolysis and glycolytic capacity; n = 3. H Representative fluorescence images of MitoSOX Red (oxidative stress marker) and DAPI (nuclear stain) staining; scale bar = 40 μm. I Quantitative analysis of MitoSOX Red-positive cells per field (30 fields counted per group). Data are presented as mean ± SD. *P ≤ 0.05, **P ≤ 0.01. Statistical analyses were performed using the methods described in Methods

To definitively establish that PDZD8 exerts neuroprotection through MERCs stabilization rather than secondary pathways, we employed the well-validated 9xL spacer construct to artificially increase ER-mitochondria distance and disrupt MERCs function [25].

Western blot analysis (Fig. 7A, B) confirmed the previously observed reduction of PDZD8 in OxyHb-treated neurons and its restoration by overexpression; importantly, co-transfection with 9xL spacer did not alter PDZD8 protein levels, excluding confounding effects on PDZD8 expression. Additional TEM analysis further verified that PDZD8 overexpression significantly shortened ER-mitochondria distance and increased ERMICC in OxyHb-treated neurons, while co-transfection with the 9xL spacer completely reversed these MERCs structural improvements, confirming the effective and specific disruption of MERCs integrity by the spacer (Fig. S4). For mitochondrial oxidative phosphorylation, OCR profiles (Fig. 7C) and quantitative analysis of basal respiration, ATP production, maximal respiration and spare respiratory capacity (Fig. 7D-G) recapitulated that OxyHb significantly impaired mitochondrial respiration while PDZD8 overexpression reversed these defects. For glycolytic metabolism, ECAR profiles (Fig. 7H) and quantitative analysis of glycolysis and glycolytic capacity (Fig. 7I-J) showed that OxyHb-induced enhancement of glycolysis was significantly inhibited by PDZD8 overexpression. Remarkably, all beneficial effects of PDZD8 overexpression on both mitochondrial and glycolytic homeostasis were completely abolished when MERCs were artificially disrupted by the 9xL spacer.

Fig. 7.

Disruption of MERCs by 9xL spacer abrogates PDZD8-mediated metabolic protection in OxyHb-treated primary neurons. A Representative Western blot images showing PDZD8 protein expression in five experimental groups. B Quantitative analysis of PDZD8 protein levels normalized to β-tubulin; n = 3. C Dynamic profiles of mitochondrial oxygen consumption rate (OCR). D–G Quantitative analysis of OCR-linked mitochondrial respiratory parameters: basal respiration, ATP production, maximal respiration, and spare respiratory capacity; n = 3. H Dynamic profiles of extracellular acidification rate (ECAR). I–J Quantitative analysis of ECAR-associated glycolytic parameters: glycolysis and glycolytic capacity; n = 3. Data are presented as mean ± SD. *P ≤ 0.05, **P ≤ 0.01, n.s. = not significant. Statistical comparisons were performed between control vs OxyHb 48 h, OxyHb 48 h+vector vs OxyHb 48 h+OE-PDZD8, and OxyHb 48 h+OE-PDZD8 vs OxyHb 48 h+OE-PDZD8+9xL spacer

Collectively, these in vitro findings indicate that PDZD8 exerts neuroprotective effects against SAH-mimicked neuronal injury by preserving MERCs structure, restoring mitochondrial metabolic balance and mitigating oxidative stress.

S-Nitrosylation at C884 promotes PDZD8 degradation via ubiquitination to mediate its post-SAH downregulation

To delineate the molecular mechanism underlying PDZD8 protein downregulation following SAH, we first assessed PDZD8 mRNA expression using qRT-PCR. No significant difference was observed between the sham and SAH 48 h groups (Fig. 8A), ruling out transcriptional regulation as the contributing factor. Building on our team’s previous observation that global S-nitrosylation levels are significantly elevated as early as 12 h post-SAH [24], we further detected total S-nitrosylation levels at 48 h post-SAH. Representative Western blot images (Fig. 8B) and quantitative analysis (Fig. 8C) confirmed a marked increase in global S-nitrosylation in the SAH group, suggesting that S-nitrosylation might be involved in the post-transcriptional regulation of PDZD8. In HT22 cells, 48 h OxyHb treatment (mimicking SAH in vitro) significantly elevated SNO-PDZD8 (Fig. 8D-E) while reducing total PDZD8 protein abundance (Fig. 8F), establishing a negative correlation between PDZD8 S-nitrosylation and its stability.

Fig. 8.

S-nitrosylation at C884 promotes PDZD8 ubiquitination and degradation to mediate its downregulation post-SAH. A Quantitative real-time PCR (qRT-PCR) analysis of PDZD8 mRNA expression in the sham and SAH 48 h groups (n=6 per group). B-C Representative Western blot images (B) and quantitative analysis (C) of global S-nitrosylation levels in the sham and SAH groups (n=6 per group). D-F HT22 cells were treated with oxyHb for 48 h to mimic SAH in vitro. Representative Western blots (D) and quantification (E) of S-nitrosylated PDZD8 (SNO-PDZD8) and total PDZD8 (F) (n=3 per group). G-H Schematic of PDZD8-knockout (PDZD8-KO) HT22 cell line construction via CRISPR/Cas9 (G) and Western blot validation of PDZD8 ablation (H). I-J PDZD8-KO cells were transfected with wild-type (WT) PDZD8 or C884A mutant plasmids (cysteine-to-arginine substitution). Representative Western blots (I) and quantification (J) of SNO-PDZD8 and total PDZD8 in control and oxyHb-treated subgroups (n=3 per group). K-L Cycloheximide (CHX) chase assay in HT22 cells stably expressing PDZD8 WT or C884A. Representative blots (K) and quantification (L) showed attenuated PDZD8 reduction in the C884A group, indicating enhanced protein stability (n=3 per group). M Ubiquitination immunoprecipitation (IP) assay showed lower ubiquitination levels of PDZD8 C884A compared to WT PDZD8 (n=3 per group). N Schematic diagram illustrating the molecular mechanism underlying PDZD8 downregulation post-SAH. Data are presented as mean ± SD. N.S., not significant, *P ≤ 0.05, **P ≤ 0.01. Statistical analyses were performed using the methods described in Methods. Figure 8G was created with BioRender.com

Subsequently, we constructed a PDZD8-knockout (PDZD8-KO) HT22 cell line using CRISPR/Cas9 gene-editing technology (schematic, Fig. 8G), and Western blot analysis validated the complete ablation of PDZD8 protein expression in the established cell line (Fig. 8H). Utilizing the deep learning-based online tool pCysMod (http://pcysmod.omicsbio.info/action.php) with PDZD8’s full-length amino acid sequence (NP_001028394.2) and “S-nitrosylation” set as the target modification, we successfully identified cysteine at position 884 (C884) as a high-confidence S-nitrosylation site (FPR ≤ 0.01, Fig. S5 and Table S7). We then constructed a C884A mutant plasmid (cysteine-to-arginine substitution) and transfected it, along with wild-type (WT) PDZD8, into PDZD8-KO cells. Western blot assays were performed to detect SNO-PDZD8 and total PDZD8 levels across eight experimental groups (four control/KO/rescue subgroups; four OxyHb-treated subgroups) (Fig. 8I), and quantitative analysis (Fig. 8J) showed that the KO + WT rescue group recapitulated the PDZD8 expression and S-nitrosylation patterns observed in WT cells.

Under OxyHb stimulation, the C884A mutation significantly reduced the SNO-PDZD8/PDZD8 ratio and increased total PDZD8 protein levels compared with the KO + WT rescue group.

To further verify whether S-nitrosylation at C884 modulates PDZD8 protein stability, we treated HT22 cells stably expressing PDZD8 WT or C884A with cycloheximide and monitored the protein half-life over a 48-hour period. Notably, the reduction in PDZD8 protein levels was significantly attenuated in the C884A group compared to the WT group (Fig. 8K-L), indicating enhanced stability of the mutant protein. To dissect the downstream mechanism underlying the enhanced stability of PDZD8 C884A, we performed ubiquitination immunoprecipitation assays. The results showed that the ubiquitination level of PDZD8 C884A was significantly lower than that of WT PDZD8 (Fig. 8M), explaining its reduced degradation rate. Collectively, these results illustrate that S-nitrosylation at the C884 residue of PDZD8 promotes its ubiquitination, thus accelerating its proteasomal degradation and reducing protein stability. In contrast, the C884A mutation blocks PDZD8 S-nitrosylation and subsequent ubiquitination, thereby alleviating its degradation and mediating PDZD8 downregulation post-SAH.

Discussion

SAH remains a devastating cerebrovascular disorder with high mortality and poor long-term prognosis, which is largely attributed to EBI occurring within 72 h post onset [27]. Despite extensive research efforts, the molecular mechanisms underlying EBI have not been fully elucidated, and there is an urgent need to identify effective therapeutic targets. In the present study, we identified PDZD8, a key regulator of MERCs, as a novel neuroprotective factor in SAH-induced EBI. Our findings demonstrated that PDZD8 is neuron-specifically downregulated at 48 h post SAH, and its overexpression alleviates EBI by preserving MERCs integrity, restoring mitochondrial homeostasis and improving neurobehavioral outcomes. Mechanistically, SAH-induced PDZD8 downregulation is mediated by S-nitrosylation at the C884 residue, which promotes PDZD8 ubiquitination and subsequent proteasomal degradation (Fig. 8N). These results uncover a previously unrecognized regulatory pathway and indicate that PDZD8 may serve as a potential therapeutic target for SAH.

MERCs, specialized membrane subdomains that tether mitochondria to the endoplasmic reticulum, play a pivotal role in sustaining cellular physiological homeostasis by mediating critical processes including calcium trafficking, lipid biosynthesis, ROS signaling transduction, and proteostasis regulation—with their dysregulation strongly linked to neurodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease, as well as stroke [16, 28, 29]. In the present study, we observed prominent structural abnormalities in MERCs at 48 h post-SAH, characterized by widened intermembrane gaps and diminished contact areas, and this pathological alteration was significantly correlated with downregulated expression of PDZD8; as an endoplasmic reticulum-resident membrane protein known to maintain MERCs stability via binding mitochondrial lipids [17], our study further confirmed that PDZD8 overexpression could efficiently reverse these impairments in both in vivo and in vitro models, expanding the current understanding of the regulatory network of MERCs in hemorrhagic brain injury.

Notably, Hao et al. reported an increase in MERCs at 24 h post-SAH and proposed that this alteration was associated with DRP1-mediated mitochondrial fission [30]. In contrast, our study identified a reduction in MERCs integrity at 48 h post-SAH accompanied by downregulation of PDZD8. One possible explanation for this discrepancy is that MERCs undergo dynamic remodeling during SAH progression and may exhibit temporal heterogeneity across different injury stages and experimental conditions. Future studies with finer temporal resolution will be valuable to clarify the dynamic relationship among MERCs remodeling, mitochondrial responses, and PDZD8 regulation following SAH.

Mitochondrial dysfunction serves as a convergence hub for multiple pathological cascades during EBI progression, where oxidative stress, energy metabolic imbalance, and calcium overload synergistically exacerbate mitochondrial impairment, thereby triggering neuronal apoptosis [8, 31]. In the present study, we observed that MERCs disruption post-SAH was accompanied by compromised mitochondrial respiratory function, compensatory upregulation of glycolysis, and accumulation of mitochondria-derived ROS, whereas PDZD8 overexpression effectively reversed these pathological alterations and restored mitochondrial metabolic homeostasis. Notably, this reparative effect was not mediated by the direct modulation of intrinsic mitochondrial quality control pathways (e.g., mitophagy, mitochondrial fission-fusion cycle, or UPRmt responses [32]) by PDZD8; instead, it relied on preserving the structural integrity of MERCs—a pivotal hub for interorganellar crosstalk—to re-establish functional synergy between mitochondria and the ER in calcium trafficking, lipid exchange, and ROS signaling transduction. To provide definitive causal evidence for this mechanistic model rather than mere correlational observations, we employed the well-validated 9xL spacer construct to artificially disrupt MERCs structure by increasing the physical distance between ER and mitochondria [25]. Our in vitro experiments demonstrated that while PDZD8 overexpression effectively reversed OxyHb-induced impairments in mitochondrial oxidative phosphorylation and glycolytic homeostasis, these beneficial effects were markedly attenuated when MERCs were artificially disrupted. These findings support the interpretation that preservation of MERCs integrity is an important component of PDZD8-mediated neuroprotection and further suggest that restoration of mitochondrial metabolic function is closely associated with MERCs maintenance.

Accumulating evidence indicates that restoration of MERCs architecture facilitates ordered calcium release from ER stores and its precise uptake by mitochondria, thus preventing mitochondrial membrane potential collapse induced by calcium overload [17, 33]; it also ensures efficient translocation of phospholipids and other lipid molecules across the two membranes, maintaining mitochondrial membrane structural stability, which in turn indirectly enhances mitochondrial respiratory chain efficiency and suppresses excessive ROS production [34–36]. Notably, our TEM analysis further corroborated this mechanism-specific effect: we observed a significant reduction in ERMICC at 48 h post-SAH, while global mitochondrial morphological parameters (perimeter and roundness index) remained unchanged. This temporal discrepancy is consistent with the hierarchical nature of subcellular injury, where disruption of inter-organelle communication precedes overt morphological damage to individual organelles [7, 37]. The fact that PDZD8 overexpression specifically restored ERMICC without altering global mitochondrial morphology provides additional strong evidence that PDZD8 exerts its neuroprotective effects primarily through stabilizing MERCs structure, rather than through a non-specific effect on mitochondrial biogenesis or dynamics.

Distinct from conventional antioxidant and anti-apoptotic therapeutic strategies, this regulatory mechanism sheds novel light on the development of targeted neuroprotective interventions against SAH.

Further investigation into PDZD8’s regulatory mechanisms revealed that its downregulation post-SAH is mediated by aberrant post-translational modifications rather than transcriptional regulation. As we have previously demonstrated that global S-nitrosylation levels are significantly elevated in the temporal cortex during the early phase post-SAH [24], this elevation is induced by excessive nitric oxide production post-SAH (primarily from iNOS and nNOS activation)—a reversible modification that contributes to various pathological processes by regulating protein stability [38, 39]. Bioinformatics prediction and site-directed mutagenesis experiments identified cysteine 884 (C884) of PDZD8 as a highly specific S-nitrosylation site, and modification at this site directly promotes PDZD8’s ubiquitination and proteasomal degradation. This mechanism shares similarities with the known S-nitrosylation regulatory pathway of Drp1²⁴, but PDZD8’s identity as a specific substrate among MERCs-regulating proteins uncovers a novel regulatory role of the S-nitrosylation-ubiquitination pathway in organelle dysfunction post-SAH, expanding understanding of SAH’s pathological molecular network. Given the global elevation of S-nitrosylation levels in the early phase post-SAH [24], it is theoretically possible that other MERCs-associated proteins may also be regulated by similar post-translational modifications. However, in the present study, we only confirmed that PDZD8 undergoes this specific cascade of site-specific S-nitrosylation at C884 followed by enhanced ubiquitination and subsequent proteasomal degradation. To the best of our knowledge, no other MERCs-specific tethering proteins have been reported to be regulated through this precise mechanism in the context of brain injury. Whether this regulatory mode is specific to PDZD8 or represents a general mechanism for the dysregulation of MERCs proteins in SAH remains an important open question that warrants further systematic investigation in future studies.

Interestingly, we observed a partial recovery of PDZD8 expression beginning at 72 h after SAH. However, although PDZD8 levels gradually increased at later time points, they remained significantly lower than those observed in sham animals throughout the observation period. At the same time, neurological deficits remained detectable despite gradual functional improvement. The biological significance of this delayed endogenous increase remains unclear. One possible explanation is that it represents a compensatory response to ongoing cellular stress after SAH. Nevertheless, whether this endogenous recovery contributes to neurological repair or is sufficient to influence long-term outcomes remains to be determined.

Further studies are required to evaluate whether delayed restoration of PDZD8 expression can promote neurological recovery after SAH.

Our statistical analyses focused on comparisons between groups under identical injury conditions to directly evaluate the effects of PDZD8 intervention while minimizing confounding introduced by SAH-induced injury. Consistent with this analytical strategy, PDZD8 overexpression under physiological conditions did not significantly affect MERCs structure or mitochondrial function. This observation may be explained by two factors. First, MERCs formation and maintenance depend on coordinated interactions among multiple tethering components, and increased expression of PDZD8 alone may not be sufficient to further enhance contact formation under physiological conditions [24]. Second, PDZD8 mainly participates in maintaining basal calcium and metabolic homeostasis under normal conditions, where these processes are already tightly regulated [17, 40]. Therefore, in contrast to normal physiological states, the functional contribution of PDZD8 may become more apparent under pathological conditions associated with MERCs disruption and mitochondrial stress.

This study has several limitations that require further refinement. First, only male experimental animals were used, while gender differences in SAH pathological processes and therapeutic responses have been confirmed by multiple studies [41]. Future research should include female animal models to clarify the gender-specific role of PDZD8. Second, we performed long-term neurological behavioral evaluation up to 21 days after SAH; nevertheless, SAH is frequently accompanied by delayed complications such as cerebral vasospasm and delayed cerebral ischemia[42]. These secondary pathological injuries may still develop and progress within the 21-day observational window, thereby affecting the stability and interpretation of long-term neurological functional outcomes. Meanwhile, whether PDZD8 participates in the occurrence and development of delayed cerebral ischemia and cerebral vascular remodeling needs further verification through long-term follow-up experiments. Third, in vitro experiments were limited to a single cell type (neurons), whereas brain injury post-SAH results from the interaction of neurons, glial cells, and vascular cells.

Subsequent studies should use cell co-culture models to explore PDZD8’s regulatory role in intercellular communication. In addition, although PDZD8 expression exhibited partial recovery at later time points after SAH, the functional significance of this endogenous response remains unclear. Future studies are needed to determine whether delayed augmentation of PDZD8 expression can influence long-term neurological recovery and tissue repair after SAH. Finally, in vivo functional verification of C884 site modification remains incomplete; constructing PDZD8 C884A knock-in animal models will further confirm the core role of this site in SAH pathology.

Conclusion

This study identifies PDZD8 as a neuron-specific regulator of MERCs in SAH-induced EBI. SAH downregulates PDZD8 via C884 S-nitrosylation-mediated ubiquitination and degradation, leading to MERCs disruption, mitochondrial dysfunction, and neuronal damage.

PDZD8 overexpression preserves MERCs integrity, restores mitochondrial homeostasis, and ameliorates neurobehavioral deficits in vivo and in vitro. Collectively, these findings highlight PDZD8 as a promising therapeutic target for mitigating EBI and improving outcomes in SAH.

Abbreviations

SAH

Subarachnoid hemorrhage

EBI

Early brain injury

ROS

Reactive oxygen species

ER

Endoplasmic reticulum

MERCs

Mitochondria-endoplasmic reticulum contact sites

PDZD8

PDZ domain containing 8

OxyHb

Oxyhemoglobin

WB

Western blotting

qRT-PCR

Quantitative reverse transcription polymerase chain reaction

TEM

Transmission electron microscopy

mNSS

Modified neurological severity scores

OFT

Open field tests

PLA

Proximity ligation assay

OCR

Oxygen consumption rate

ECAR

Extracellular acidification rate

SNO

S-nitrosylated

ERMICC

ER-mitochondria contact coefficient

Authors' contributions

Peng Deng: Conceptualization; Data curation; Formal analysis; Writing – original draft Xin Wu: Data curation; Formal analysis; Writing – original draft Yangyang Liu: Data curation; Formal analysis; Writing – original draft Bingyi Song: Investigation; Methodology; Project administration Xi Liu: Investigation; Methodology; Project administration Xiao Yang: Methodology; Project administration Han Luo: Project administration Jiabei Peng: Resources; Software Haiying Li: Supervision; Validation; Writing – review & editing Zhong Wang: Supervision; Validation Zongqi Wang: Conceptualization; Funding acquisition; Supervision; Writing – review & editing Wen Li: Conceptualization; Writing – review & editing.

Funding

This work was supported by the Gusu Talent Program (GSWS2023022).

Data availability

Data will be made available on request.

Declarations

Ethics approval and consent to participate

Animal experiments followed the guidelines of the First Affiliated Hospital of Suzhou University, approved by the Ethics Committee (GSWS2023022).

Consent for publication

All authors confirm reading and approving the manuscript, taking full accountability for its accuracy and integrity. The article is original, unpublished, and not under review elsewhere.

Competing interests

The authors declare no competing interests.